‘Questions become answers in seconds’: AI’s real impact on lending

Most of the discussion around AI in specialist finance focuses on what it could do in the future.

To a degree, that misses the point. The real question is what practical steps businesses can take today to use AI safely, productively and at scale.

In a recent Lunch & Learn, I told the TAB team that the biggest AI shift is the way people interact with data. AI turns analysis into a conversation. Instead of defining every requirement upfront, you can move away from lengthy specification documents and towards a more iterative approach. Ask a question, review the answer, refine your thinking and ask another.

That changes the speed at which decisions are made, too. For years, analytics followed a familiar pattern. Someone asked a question. A data team built a report. A dashboard appeared a few weeks later. By the time the answer arrived, the business had often moved on. Today, AI is shifting analytics away from static reporting. Instead of waiting for reports, teams can interrogate live datasets directly.

Questions become answers in seconds, not days. You can upload spreadsheets, documents and written queries into AI tools and generate working outputs almost instantly. Non-technical teams can analyse datasets that previously required specialist technical support. The gap between question and answer has narrowed dramatically. When you’re trying to deliver commercial mortgages at bridging speed, that matters.

At TAB, we see practical applications every day.

Take our investors. We fund our commercial mortgages and bridging finance through a combination of institutional capital and funding from private investors. We can now analyse investor behaviour across different asset classes, identifying emerging trends in borrower demand. We can examine regional concentrations in the loan book. We can explore repayment patterns and refinancing behaviour. We can test assumptions against real data rather than instinct.

Risk concentrations can be identified without lengthy development projects. Portfolio trends become visible earlier. Business leaders gain direct access to information rather than relying exclusively on specialist teams.

I offer a word of warning to business leaders within the industry starting out on the AI journey. One of the big AI risks is not the technology itself. It is confidence in answers that rest on inconsistent data definitions or flawed assumptions. If two employees ask similar questions and receive different answers, confidence evaporates quickly. Different users generate potentially unreliable interpretations unless there is a shared layer of definitions and control over underlying data logic.

That’s just one of the reasons why governance is becoming more important with the adoption of AI.

We have put acceptable use policies in place as well and added AI to our risk register and business continuity plans. We have strengthened our internal controls and we continue to invest in security. The conversation around AI often focuses on productivity gains while the cybersecurity implications receive less attention.

Powerful AI models have the capability to find previously hidden IT vulnerabilities. They could potentially be used in hacking. The Bank of England’s Prudential Regulation Authority is understandably concerned about the vulnerabilities in lenders’ IT systems being exposed by the latest AI models, warning the technology could soon fall into the wrong hands.

The reality is simple. As AI becomes more powerful, the consequences of weak controls become more severe. That is why responsible adoption counts. Not because regulation demands it. Because good businesses demand it.

The next phase of AI adoption in commercial mortgages and bridging finance will not be defined by grand transformation programmes or theoretical debates. It will be defined by practical use cases. Teams sharing what works and people solving real operational problems.

The organisations that make the most progress won’t necessarily be the largest. They will be the most curious and the most consistent in how they test, apply and refine what AI shows them. The advantage will sit with those that move from isolated experiments to repeatable ways of working.